Hi folks,
I am having a strange performance problem and I am wondering if anyone has seen
something similar.
I have 2 8400's with 8g of RAM, 10 612MHz processors, 9g swap using TruCluster
Available server on DU4.0d. Here is the scoop. I can run through some screens
in my DB as quick as can be and then boom, I will get a 10-20 second delay. Mind
you, it would not coinside with any sort of DB lookup/update, this will arise
once the data had been retrieved and is in memory. This will happen when the
system is busy as well as when it is quiet (i.e. low disk I/O, CPU's 80+ idle,
low cs and sy,). Swapping does not occur. This slowdown did not coinside with
any sort of kernel change or patch. I have made several improvements and nothing
seems to have an impact on this intermittant freeze.
I can try this against the same piece of code on our development system (same
type of system, same OS level, same patch level) and I do not see this problem.
And this system has only 8 cpu's, with less controllers and much more congested
disks.
Basically, I am starting to wonder if there is some sort of hardware problem
that is not getting recorded. Has anyone out there had a system with similar
symptoms?
Thanks so much.
-cherie
Received on Thu Oct 07 1999 - 22:05:48 NZDT